USGS - science for a changing world

Northern Prairie Wildlife Research Center

  Home About NPWRC Our Science Staff Employment Contacts Common Questions About the Site

Impact of Red Fox Predation on the Sex Ratio of Prairie Mallards


While we are confident of the general validity of our results, we offer several cautions in their interpretation. Mostly, these involve matters where data used in our model were based on unreplicated studies or were completely lacking and we had to make subjective appraisals. In many cases, subjects worthy of additional research effort are indicated.

One topic meriting further investigation is the nature and extent of predation by foxes. We have used the best available data, and opted for the conservative figure when confronted with a choice. Yet, the expansion factors are considerable and we would feel more comfortable if additional information existed on, in particular, the proportion of mallards taken by foxes but not brought to the dens. We relied on but one study for our estimate of this proportion. The other components of the predation rate also deserve more study, especially the number of dens used by prairie red foxes, but our estimates are based on broader investigations. Little is known about the densitydependence of fox predation rates. Our initial model incorporated an upper bound on the rate, the value of which, despite its arbitrariness, did not markedly affect the output of the model, as demonstrated by our sensitivity analysis. The model in Part Two incorporated a density-dependency relationship that was realistic in form, but still derived subjectively. Detailed field studies would be required to furnish information on the true relationship.

An open question in mallard population dynamics involves hunting mortality, its measurement, and its role. Assessing the rate of hunting mortality by using recoveries of banded birds depends markedly upon the reporting rate and, to a lesser extent, the nonretrieval of birds. The reporting rate is known to fluctuate widely according to location of recovery, distance from banding site, attitudes of hunters, and numbers of birds banded. All of these factors influence the estimated rate of hunting kill. More important perhaps than estimating hunting mortality is determining how it fits into total annual mortality. We have followed the usual assumption that hunting represents an additive form of mortality, so in our model fox-related and "other" mortality rates do not depend upon the hunting rates (although actual numbers do, of course, because casualties of one source of mortality are unavailable for another). Anderson and Burnham (1976) recently questioned this model and argued that hunting mortality, up to a certain level, was largely compensatory to other forms of mortality in mallards. Because their theory has not yet been fully evaluated, and because they suggested no mechanism which might compensate for hunting mortality, we chose to stay with the conventional assumption that it is additive.

The losses of mallards to foxes and to hunters are relatively well known in comparison to losses to other predators and to nonpredation causes. Qualitative knowledge is available, e.g., that eagles kill some mallards, but quantitative data are lacking. Even in instances where total mortality is known, e.g., an outbreak of botulism that killed 10,000 mallards, we cannot fathom the impact on the total population, as compared to less spectacular but more regular causes of death. Fortunately, our discussion of other mortality factors did not depend strongly upon their magnitudes, but mostly on their relative effects on the sexes.

The primary output of our model, the simulated spring sex ratio of mallards was compared to actual field counts. As has been emphasized by several waterfowl researchers, however, the accuracy of those counts is questionable, as they vary by phenology of the nesting season, weather conditions, and location. More work needs to be done to ascertain the validity and variability of spring waterfowl counts and to determine optimal census methods for obtaining sex ratios.

Notwithstanding any shortcomings in the data, we conclude that the sex ratio of the North Dakota mallard population could have changed from 100:100 in 1963 to about 126:100 in 1973, as a result of depressed survival of adult females compared to males. More significantly, the reduced female survival can be attributed largely to predation (including scavenging) by red foxes. The model in Part Two gave an average sex ratio of about 118:100, a more representative value reflecting conditions throughout the period. Scrutiny of other causes of mortality affecting mallards did not reveal any that were sufficiently hen-selective and intense to have substantial effect on the sex ratio. Hunting was the principal male-selective mortality factor, which partially restored sexual balance.

The model proved to be both robust and stable and yielded results that were validated by field data of several kinds, including sex ratios of actual prenesting counts of mallards. Certain facets of the results were further verified by direct mathematical techniques. Thus, we conclude with a reasonable degree of certainty that red fox predation (including scavenging of birds) alone is adequate to distort the sex ratio of a mallard population to the extent observed in the field.

Modeling provided the vehicle whereby the above conclusions could be reached and clearly identified those factors that had the greatest effect on the sex ratio. The model also provided the means to make predictive estimates with management overtones and to explore the nature of pristine sex ratios. This, of course, was possible only because we possessed adequate data on the necessary parameters of the model.

Mallard sex ratios were found to be sensitive to fox densities, hunting rates, and mallard densities. A reduced fox population caused nearly even sex ratios, and a higher fox population resulted in further distortion. Hunting, too, exerted a strong influence on the sex ratio; a relatively small increase in male mortality with no change in female mortality counteracted the effect of foxes and resulted in a more even ratio. The mallard population level was not proportionately as important as the other two variables although a doubling of the population would nearly achieve sexual balance. Of the three variables, hunting is the most easily manipulated but, because of its potency, it must be carefully employed.

Historical data led us to conclude that the sex ratio of mallards in the Prairie Pothole Region has become increasingly a function of human-influenced mortality factors and that the impact of red foxes has increased dramatically during the past 40 years. Red foxes had little effect on mallard sex ratios during pristine times and, even though other canids were present, their combined influence was likely less than that of foxes today. Red foxes are now the major single source of nonhunting mortality on hens in the Prairie Pothole Region. While foxes probably exert their greatest influence in North Dakota, they have become more important in other parts of the region in recent years. The situation has not stabilized, however, as canid populations in much of the prairie are still in a state of flux.

Although management attention has focused primarily on the possibility of harvesting the "excess" drakes, the role of these drakes in breeding populations has not been evaluated. The notion that populations have evolved with a slight surplus of drakes suggests that the presence of at least some "extra" drakes serves a useful function. Until data are available to define the role of drakes in populations of various sex compositions, we believe that it would be inadvisable to strive through management for a sex ratio below about 110:100, the probable pristine figure, and one which includes a small surplus of drakes in case they do contribute to the productivity.

If our interpretation is accurate, the "excess" drakes are not truly extra but rather a result of a relatively new additional loss of hens that has developed since pristine times. Although it may be reasonable to harvest those drakes, it might be best to do so with the thought of sparing the hens rather than harvesting a surplus component of the population.

Distorted sex ratios occur among most prairie-dwelling ducks, but the greatest imbalance seems to be with certain diving ducks, in which red fox predation is of little consequence. We suggest that greater sexual imbalance among canvasbacks, for example, probably occurs because important mortality factors have compounding rather than compensating effects on the sex ratio. Predation (mostly by mink) is probably hen-selective. Hunting also seems to be hen-selective, a situation opposite that in mallards.

Although mallard sex ratios could be managed by altering fox populations or modifying the productivity of the bird, the approach usually taken has been through hunting regulations. The sex ratio is but one of several factors responding to changes in hunting intensity, and perhaps less important than other aspects of the population dynamics, particularly total mortality. Although the primary management implications of a disproportionate sex ratio relate to hunting, there are important implications for the interpretation of spring breeding pair counts. If sex ratios of mallards or other ducks are as distorted as suggested by this study, then, depending on the behavior of the supernumerary drakes, the very real possibility exists that spring estimates of breeding pairs are overestimated because they are derived from counts of drakes under the assumption that all drakes seen in groups of one to five represent pairs.

Previous Section -- Management Implications
Return to Contents
Next Section -- Acknowledgments

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo logo U.S. Department of the Interior | U.S. Geological Survey
Page Contact Information: Webmaster
Page Last Modified: Saturday, 02-Feb-2013 05:55:44 EST
Reston, VA [vaww55]